Abstract: System, methods, and computer-readable media for configuring an autonomous vehicle based on safety scores determined by a safety score prediction algorithm, and an associated training technique, to output a safety score for a predicted and/or actual path. The safety score is a value indicating a likelihood of risky events per distance. The safety score prediction algorithm is trained with historical human driving datasets associated with paths (predicted or actual) taken by one or more AVs during human driving.

Abstract: The present technology is directed to training and the use of a machine learning model to measure the safety of an autonomous vehicle (AV) driving in simulation. An AV management system can run a simulation of an AV autonomously piloting itself and collect simulation driving data. Further, the AV management system can parse the simulation driving data into kinematic and semantic environmental features and output a simulation safety score of the simulation based on the kinematic and semantic environmental features. The simulation safety score indicates a probability of a safety critical event such as a collision or a near-miss between the AV and the at least one simulated object in the simulation.

Abstract: Various technologies described herein pertain to routing autonomous vehicles based upon spatiotemporal factors. A computing system receives an origin location and a destination location of an autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a spatiotemporal statistical model. The spatiotemporal statistical model is generated based upon historical data from autonomous vehicles when the autonomous vehicles undergo operation-influencing events. The spatiotemporal statistical model takes, as input, a location, a time, and a direction of travel of the autonomous vehicle. The spatiotemporal statistical model outputs a score that is indicative of a likelihood that the autonomous vehicle will undergo an operation-influencing event due to the autonomous vehicle encountering a spatiotemporal factor along a candidate route.

Abstract: Various technologies described herein pertain to routing an autonomous vehicle based upon risk of takeover of the autonomous vehicle by a human operator. A computing system receives an origin location and a destination location of the autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a computer-implemented model. The computer-implemented model is generated based upon labeled data indicative of instances in which autonomous vehicles are observed to transition from operating autonomously to operating based upon conduction by human operators while the autonomous vehicles are executing predefined maneuvers. The computer-implemented model takes, as input, an indication of a maneuver in the predefined maneuvers that is performed by the autonomous vehicle when the autonomous vehicle follows a candidate route.

Abstract: Described herein are technologies relating to computing a likelihood of an operation-influencing event with respect to an autonomous vehicle at a geographic location. The likelihood of the operation-influencing event is computed based upon a prediction of a value that indicates whether, through a causal process, the operation-influencing event is expected to occur. The causal process is identified by means of a model, which relates spatiotemporal factors and the operation-influencing events.

Abstract: System, methods, and computer-readable media for configuring an autonomous vehicle based on safety scores determined by a safety score prediction algorithm, and an associated training technique, to output a safety score for a predicted and/or actual path. The safety score is a value indicating a likelihood of risky events per distance. The safety score prediction algorithm is trained with historical human driving datasets associated with paths (predicted or actual) taken by one or more AVs during human driving.

Abstract: The present technology is directed to training and using a machine learning model to predict a likelihood of a counterfactual safety critical event in autonomous vehicle (AV) driving in a projected scenario occurring after a human takes over control of an AV. An AV management system can identify driving data collected from periods around an occurrence of a human take over event where a human takes over control of an AV. The AV management system can project a scenario that would have resulted if the human did not take over control of the AV based on the driving data and output a counterfactual safety score for the projected scenario. The counterfactual safety score can indicate a probability of a counterfactual collision between the AV and the at least one object in the projected scenario.

Abstract: The present technology is directed to training and the use of a machine learning model to measure the safety of an autonomous vehicle (AV) driving in simulation. An AV management system can run a simulation of an AV autonomously piloting itself and collect simulation driving data. Further, the AV management system can parse the simulation driving data into kinematic and semantic environmental features and output a simulation safety score of the simulation based on the kinematic and semantic environmental features. The simulation safety score indicates a probability of a safety critical event such as a collision or a near-miss between the AV and the at least one simulated object in the simulation.

Abstract: The present technology is directed to training and the use of a machine learning model to measure the safety of an autonomous vehicle (AV) driving. An AV management system can identify driving data including sensor data from an AV that is descriptive of an environment around the AV, a path of the AV, kinematic data of the AV, a path of at least one object in the environment, and in-memory data pertaining to data output by one or more algorithms in an autonomous driving stack. As follows, the AV management system can output a safety score for the path of the AV indicating a probability of a collision between the AV and the at least one object.

Abstract: Described herein are technologies relating to computing a likelihood of an operation-influencing event with respect to an autonomous vehicle at a geographic location. The likelihood of the operation-influencing event is computed based upon a prediction of a value that indicates whether, through a causal process, the operation-influencing event is expected to occur. The causal process is identified by means of a model, which relates spatiotemporal factors and the operation-influencing events.

Abstract: Various technologies described herein pertain to routing autonomous vehicles based upon spatiotemporal factors. A computing system receives an origin location and a destination location of an autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a spatiotemporal statistical model. The spatiotemporal statistical model is generated based upon historical data from autonomous vehicles when the autonomous vehicles undergo operation-influencing events. The spatiotemporal statistical model takes, as input, a location, a time, and a direction of travel of the autonomous vehicle. The spatiotemporal statistical model outputs a score that is indicative of a likelihood that the autonomous vehicle will undergo an operation-influencing event due to the autonomous vehicle encountering a spatiotemporal factor along a candidate route.

Abstract: Described herein are technologies relating to computing a likelihood of an operation-influencing event with respect to an autonomous vehicle at a geographic location. The likelihood of the operation-influencing event is computed based upon a prediction of a value that indicates whether, through a causal process, the operation-influencing event is expected to occur. The causal process is identified by means of a model, which relates spatiotemporal factors and the operation-influencing events.

Abstract: Various technologies described herein pertain to routing autonomous vehicles based upon spatiotemporal factors. A computing system receives an origin location and a destination location of an autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a spatiotemporal statistical model. The spatiotemporal statistical model is generated based upon historical data from autonomous vehicles when the autonomous vehicles undergo operation-influencing events. The spatiotemporal statistical model takes, as input, a location, a time, and a direction of travel of the autonomous vehicle. The spatiotemporal statistical model outputs a score that is indicative of a likelihood that the autonomous vehicle will undergo an operation-influencing event due to the autonomous vehicle encountering a spatiotemporal factor along a candidate route.

Abstract: Various technologies described herein pertain to routing an autonomous vehicle based upon risk of takeover of the autonomous vehicle by a human operator. A computing system receives an origin location and a destination location of the autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a computer-implemented model. The computer-implemented model is generated based upon labeled data indicative of instances in which autonomous vehicles are observed to transition from operating autonomously to operating based upon conduction by human operators while the autonomous vehicles are executing predefined maneuvers. The computer-implemented model takes, as input, an indication of a maneuver in the predefined maneuvers that is performed by the autonomous vehicle when the autonomous vehicle follows a candidate route.

Abstract: Described herein are technologies relating to computing a likelihood of an operation-influencing event with respect to an autonomous vehicle at a geographic location. The likelihood of the operation-influencing event is computed based upon a prediction of a value that indicates whether, through a causal process, the operation-influencing event is expected to occur. The causal process is identified by means of a model, which relates spatiotemporal factors and the operation-influencing events.

Abstract: Various technologies described herein pertain to routing an autonomous vehicle based upon risk of takeover of the autonomous vehicle by a human operator. A computing system receives an origin location and a destination location of the autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a computer-implemented model. The computer-implemented model is generated based upon labeled data indicative of instances in which autonomous vehicles are observed to transition from operating autonomously to operating based upon conduction by human operators while the autonomous vehicles are executing predefined maneuvers. The computer-implemented model takes, as input, an indication of a maneuver in the predefined maneuvers that is performed by the autonomous vehicle when the autonomous vehicle follows a candidate route.

Abstract: Described herein are technologies relating to computing a likelihood of an operation-influencing event with respect to an autonomous vehicle at a geographic location. The likelihood of the operation-influencing event is computed based upon a prediction of a value that indicates whether, through a causal process, the operation-influencing event is expected to occur. The causal process is identified by means of a model, which relates spatiotemporal factors and the operation-influencing events.

Abstract: Described herein are technologies relating to computing a likelihood of an operation-influencing event with respect to an autonomous vehicle at a geographic location. The likelihood of the operation-influencing event is computed based upon a prediction of a value that indicates whether, through a causal process, the operation-influencing event is expected to occur. The causal process is identified by means of a model, which relates spatiotemporal factors and the operation-influencing events.

Abstract: Various technologies described herein pertain to routing autonomous vehicles based upon spatiotemporal factors. A computing system receives an origin location and a destination location of an autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a spatiotemporal statistical model. The spatiotemporal statistical model is generated based upon historical data from autonomous vehicles when the autonomous vehicles undergo operation-influencing events. The spatiotemporal statistical model takes, as input, a location, a time, and a direction of travel of the autonomous vehicle. The spatiotemporal statistical model outputs a score that is indicative of a likelihood that the autonomous vehicle will undergo an operation-influencing event due to the autonomous vehicle encountering a spatiotemporal factor along a candidate route.

Abstract: Various technologies described herein pertain to routing an autonomous vehicle based upon risk of takeover of the autonomous vehicle by a human operator. A computing system receives an origin location and a destination location of the autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a computer-implemented model. The computer-implemented model is generated based upon labeled data indicative of instances in which autonomous vehicles are observed to transition from operating autonomously to operating based upon conduction by human operators while the autonomous vehicles are executing predefined maneuvers. The computer-implemented model takes, as input, an indication of a maneuver in the predefined maneuvers that is performed by the autonomous vehicle when the autonomous vehicle follows a candidate route.